Since most large AI models are becoming similar, avoid betting on a single company to create a "winner-take-all" model. Instead, focus on the "picks and shovels" of the AI revolution, which are essential for the entire industry's growth. Invest in companies that provide the core hardware, such as semiconductors, and those that own the cloud infrastructure where AI models are trained. Another strong strategy is to identify companies that excel at integrating various AI models into useful products and services. For broad, diversified exposure to the entire sector, consider an AI-focused ETF.

Artificial Intelligence (AI) Sector

• The discussion centers on the idea that most large AI models are becoming interoperable and very similar in capability. The speaker suggests that any competitive edge one model has over another is marginal and temporary.
• The core reason for this convergence is that all models are learning from the same vast pool of data on the internet, and they are also learning from each other's outputs.
• The speaker views Artificial General Intelligence (AGI) not as a single product created by one company, but as a "state of being" that will emerge across the entire interconnected AI network. It will start somewhere but will quickly spread everywhere.

Takeaways

• Avoid betting on a single "winner": The transcript suggests that picking one company based on the belief that it will create the single best AI model may be a flawed strategy. Since the models are all becoming similar, a "winner-take-all" scenario in the AI model space is unlikely.
• Focus on the broader ecosystem: A more robust investment strategy might be to focus on the "picks and shovels" of the AI revolution. This includes:
  • Companies that provide the essential hardware (like semiconductors and chips).
  • Companies that own the cloud infrastructure where these models are trained and run.
  • Companies that are best at integrating various AI models into useful products and services for consumers and businesses.
• Consider diversification: Given the difficulty in picking a single winner, investors might consider a diversified approach, such as an AI-focused ETF. This would provide exposure to the growth of the entire sector without the risk of being tied to a single company whose model may become commoditized.
• Think long-term: The concept of AGI as an emergent "state" implies a long-term technological shift. The investment thesis here is not about short-term outperformance of one model, but about the long-term, transformative power of the entire AI network.